Video summary

How to Send Cold DMs on Instagram With AI Automation

Main summary

Key takeaways

Business

Core strategy / problem framing

  • Cold DMs can generate clients even with a limited budget, but manual execution becomes a bottleneck (“no one wants to sit there sending DMS all day”).
  • The key constraint is that Instagram is strict about automation—most approaches risk bans or restrictions.
  • The “play” presented is:
    • Use an automation tool designed to behave like real user activity (via mobile proxies).
    • Automate the sales conversation after the DM is received.

Tool + operating model (Auto IGDM)

Auto IGDM is positioned as the only approach the presenter trusts for safe Instagram DM automation.

How they claim it stays safe / reliable

  • Real mobile proxies per account, including:
    • Mobile-proxy cost is high
    • Hard to replicate
  • Warming up behavior before sending:
    • Accounts perform small initial actions instead of instantly blasting DMs
  • Emphasis: not for “DM blasts from hundreds of accounts,” but for targeted outreach that lands in inboxes.

Execution workflow

  1. Sync/authorize the Instagram account using a security key / authenticator secret key (presented as mandatory for reliable syncing without flags).
  2. Send outreach via campaigns:
    • Choose who to message and what to say.
  3. Use sequences:
    • AI-generated message variations to avoid repetition flags
    • Optional follow-up after a few days
  4. Use statistics to confirm throughput until the lead list runs out.

Playbook: Campaign setup & lead targeting

Lead selection is treated as the most important lever because DM volume is capped.

1) Lead source / selection method

  • Prefer people who engaged with other accounts in the niche.
    • Rationale: engagement implies activity.
  • Hashtag targeting is available but treated as less reliable due to:
    • Irrelevant hashtags
    • Users stopping effective hashtag use

2) Account selection logic

  • Use high-engagement accounts rather than huge-but-sparse accounts:
    • Avoid “millions of followers with hidden likes/comments”
    • Target accounts where engagement is visible and consistent
  • Example “niche hero” approach:
    • Target people engaging with Alex Hormozi–type verified accounts (handles mentioned in subtitles; exact spelling unclear in the transcript).

3) Hard constraints / recommended settings (interpreted from subtitles)

  • DM daily cap: ~50 DMs/day (tool/platform limitation)
    • Implication: every DM must be high-fit
  • Recommended filter defaults:
    • Min followers: typically 500 (increase for influencers)
    • Max followers: about 1,000,000 (very large accounts may ignore DMs due to volume)
    • Account activity: around 20 posts (proxy for legitimacy/active status)
  • Message both:
    • Personal and business accounts
  • Profile requirements (as described):
    • Profile picture
    • Bio/business indicators
    • Prefer accounts with a website/link in bio

Lead qualification filters: Keyword include/exclude

Keyword filtering against bios is used to maximize relevance.

Positive keywords (examples)

  • “CEO”
  • “Millions”
  • “scale”
  • “founder”
  • “7-figure/7 scale”
  • “Brands”

(Exact list described as downloadable via a link.)

Negative keywords (examples)

  • “artists”
  • “rappers”
  • “young”
  • “music”

(Framing: these lead types are not the target.)

What the tool does with these

  • It analyzes bios of people who engaged with chosen accounts.
  • It:
    • Includes leads containing positive terms
    • Excludes leads containing negative terms

Message strategy: avoid spam + improve deliverability

Outreach message construction

  • Messages should be short and non-spammy because Instagram filters spam effectively.
  • Example DM (mentors collective campaign):
    • “what’s up Jay do you ever do podcasts if not are you interested in doing more”
  • No links or images in the initial DM (per presenter).

Variation system (anti-duplication)

  • Critical rule: Instagram may flag accounts sending the same message repeatedly.
  • Auto IGDM generates unique variations per contact using AI to reduce repetition risk.

Follow-up process (within the DM cap)

  • Follow-ups are optional and controlled by the tool.
  • Timing:
    • 2–3 days after the initial message (presenter preference)
  • Follow-up counts toward the same 50 DMs/day limit.

When to use follow-up

  • If the lead list is narrow (e.g., ~1,000 prospects), follow-up helps increase output from limited daily capacity.

Conversation automation after the reply

The second phase routes inbound DM conversations to AI or a team with minimal operational overhead.

Option A: ManyChat Pro + VA access (“team” solution)

  • ManyChat Pro costs about $15/month per account (per subtitles).
  • Approach:
    • Invite BDRs as VA users to see/respond.
  • Integration assumption:
    • Each Auto IGDM account has ManyChat Pro attached so the sales team can manage conversations without logging into Instagram directly.

Option B: GoHighLevel (GHL) + Instagram DM exchange

  • If you’re not managing many accounts:
    • Use GHL to connect Instagram and exchange DMs through one system.

Option C: AI agent inside ManyChat

  • ManyChat AI can be enabled “with a click.”
  • Logic described:
    • Trigger: AI when the user sends a message
    • Action: AI generates a response
    • Goal: qualify via 1–3 qualifying questions and book a call using a link
  • Prompting support:
    • Use ChatGPT “Prompt Engineer” to create the AI prompt for ManyChat.

Pricing / unit economics (as stated)

Auto IGDM Pro

  • $99/month per account (presented as the cloud-based/pro tier)
  • Reasoning:
    • Mobile proxies drive massive computing costs
  • Presenter’s stance:
    • “Only solution right now” that covers automation + first-name extraction + safe account behavior.

Other cost mentioned

  • ManyChat Pro: ~$15/month per account (if used for conversation handling)

Results & KPI-like claims

The presenter provides operational statistics (not formal CAC/LTV/revenue).

DM throughput

  • In statistics across several months, campaigns sent:
    • At least 50 messages per day on most days
  • Delivery stops when:
    • The lead list runs out (refresh/update the target list)

Actionable recommendations (condensed playbook)

  • Prioritize lead quality because you’re limited to ~50 DMs/day.
  • Use “engaged-with-niche-account” targeting over hashtags.
  • Filter leads using:
    • Positive bio keywords (CEO/founder/scale/millions/etc.)
    • Negative bio keywords (artists/rappers/music/young/etc.)
  • Keep first DMs:
    • Short, direct, no links/images, no spammy phrasing
  • Use AI message variations to avoid repetition flags.
  • Follow up after 2–3 days (respecting the daily DM cap).
  • Automate inbound handling via:
    • ManyChat Pro (team/VA access) or GHL
    • Or a ManyChat AI agent to qualify (1–3 questions) and book calls

Presenters / sources

  • Presenter (main): Dr. J Feldman
    • Referenced as an example account being added/synced (presenter appears to use their own team/accounts).
  • Tool/company mentioned:
    • Auto IGDM
    • ManyChat
    • GoHighLevel (GHL)
  • External figures used as examples for targeting:
    • Alex Hormozi (and “Alex rosi” as a subtitle variant)
    • “Her oos” / “Herios” (context indicates “niche hero” accounts; exact handle unclear from subtitles)

Original video